A Novel Cross-Disciplinary Multi-Institute Approach to

advertisement
ORIGINAL RESEARCH
A Novel Cross-Disciplinary Multi-Institute Approach to
Translational Cancer Research: Lessons Learned from
Pennsylvania Cancer Alliance Bioinformatics Consortium
(PCABC)
Ashokkumar A. Patel 1, John R. Gilbertson 1, Louise C. Showe 2, Jack W. London 3,
Eric Ross 4 , Michael F. Ochs 4 , Joseph Carver 5 , Andrea Lazarus 6 , Anil V. Parwani 1 ,
Rajiv Dhir 1 , J. Robert Beck 4 , Michael Liebman 7 , Fernando U. Garcia 8 , Jeff Prichard 9 ,
Myra Wilkerson 9 , Ronald B. Herberman 1 , Michael J. Becich 1 , and the PCABC.
1
Center for Pathology Informatics, Benedum Oncology Informatics Center, University of Pittsburgh
Cancer Institute 2 The Wistar Institute 3 Kimmel Cancer Center of Thomas Jefferson University 4 Fox
Chase Cancer Center 5 Abramson Cancer Center of the University of Pennsylvania 6 Pennsylvania
State Cancer Institute at Milton S. Hershey Medical Center 7 Windber Research Institute 8 Drexel
University College of Medicine 9 Geisinger Health System
Abstract
Background: The Pennsylvania Cancer Alliance Bioinformatics Consortium (PCABC, http://www.pcabc.upmc.edu) is one
of the first major project-based initiatives stemming from the Pennsylvania Cancer Alliance that was funded for four years
by the Department of Health of the Commonwealth of Pennsylvania. The objective of this was to initiate a prototype biorepository and bioinformatics infrastructure with a robust data warehouse by developing a statewide data model (1) for
bioinformatics and a repository of serum and tissue samples; (2) a data model for biomarker data storage; and (3) a public
access website for disseminating research results and bioinformatics tools. The members of the Consortium cooperate
closely, exploring the opportunity for sharing clinical, genomic and other bioinformatics data on patient samples in oncology, for the purpose of developing collaborative research programs across cancer research institutions in Pennsylvania. The
Consortium’s intention was to establish a virtual repository of many clinical specimens residing in various centers across
the state, in order to make them available for research. One of our primary goals was to facilitate the identification of cancerspecific biomarkers and encourage collaborative research efforts among the participating centers.
Methods: The PCABC has developed unique partnerships so that every region of the state can effectively contribute and
participate. It includes over 80 individuals from 14 organizations, and plans to expand to partners outside the State. This
has created a network of researchers, clinicians, bioinformaticians, cancer registrars, program directors, and executives from
academic and community health systems, as well as external corporate partners - all working together to accomplish a common mission.
The various sub-committees have developed a common IRB protocol template, common data elements for standardizing
data collections for three organ sites, intellectual property/tech transfer agreements, and material transfer agreements that
have been approved by each of the member institutions. This was the foundational work that has led to the development of
a centralized data warehouse that has met each of the institutions’ IRB/HIPAA standards.
Results: Currently, this “virtual biorepository” has over 58,000 annotated samples from 11,467 cancer patients available
for research purposes. The clinical annotation of tissue samples is either done manually over the internet or semiautomated batch modes through mapping of local data elements with PCABC common data elements. The database currently holds information on 7188 cases (associated with 9278 specimens and 46,666 annotated blocks and blood samples)
of prostate cancer, 2736 cases (associated with 3796 specimens and 9336 annotated blocks and blood samples) of breast
cancer and 1543 cases (including 1334 specimens and 2671 annotated blocks and blood samples) of melanoma. These
numbers continue to grow, and plans to integrate new tumor sites are in progress. Furthermore, the group has also developed
a central web-based tool that allows investigators to share their translational (genomics/proteomics) experiment data on
research evaluating potential biomarkers via a central location on the Consortium’s web site.
Conclusions: The technological achievements and the statewide informatics infrastructure that have been established by
the Consortium will enable robust and efficient studies of biomarkers and their relevance to the clinical course of cancer.
Correspondence: Ashokkumar A. Patel and Michael J. Becich. Email: ashok.patel@case.edu;
becich@pitt.edu
Please note that this article may not be used for commercial purposes. For further information please refer to the copyright
statement at http://www.la-press.com/copyright.htm
Cancer Informatics 2007:3 255–274
255
Patel et al
Studies resulting from the creation of the Consortium may
allow for better classification of cancer types, more accurate
assessment of disease prognosis, a better ability to identify
the most appropriate individuals for clinical trial participation, and better surrogate markers of disease progression
and/or response to therapy.
Background
Biomedical researchers, geneticists, pathologists
and tissue engineers depend on cancer biorepositories for the development and validation of new
diagnostic or prognostic cancer markers, the understanding of basic disease mechanisms, and the
evaluation of proposed therapeutic regimens or the
creation of new bio-engineered materials or pharmaceuticals. Recent advances in translational
research have resulted in a growing demand for
specific, highly annotated human tumor tissues
[1–3], and this demand has reinforced the importance of tissue banks as a major part of the necessary infrastructure of any region seeking to become
a force in medical biotechnology.
Although virtually all major academic cancer
centers—in Pennsylvania and across the nation—
have research tissue banks, it is still difficult for
most cancer researchers to find sufficient and well
annotated tumor tissues that are necessary to begin
and complete their work. This is because existing
banks are beset by a number of major limitations:
(1) the number of tissue specimens at a given site
is limited, (2) sharing of information is constrained
by confidentiality and de-identification issues, (3)
the annotation process is complex and expensive,
and (4) each tissue bank may define and apply
annotation differently. The result is that access to
well documented tissue specimens, using normalized descriptors, remains one of the most important
limitations to cancer biomarker research[4].
In 1998, Pennsylvania’s leading cancer centers
organized the Pennsylvania Cancer Alliance (PCA)
and set as one of its first goals the task of communicating to Commonwealth legislators and the
Governor the need to allocate on a long-term basis
a substantial portion of the funds from that year’s
national Master Settlement Agreement with the
tobacco industry[5] to research and prevention of
diseases directly linked to tobacco use, including
cancer. In 2001, the PCA’s efforts led to the
Commonwealth legislators’ and the Governor’s
decision to approve Act 77 [6], to allocate 19% of
Pennsylvania’s tobacco settlement money for
biomedical research initiatives that would be
256
sponsored by the PA Department of Health’s
Commonwealth Universal Research Enhancement
(CURE) program [7]. The details of the allocated
funds to this project are discussed below.
The Pennsylvania Cancer Alliance Bioinformatics Consortium (PCABC, http://www.pcabc.
upmc.edu) is one of the first major project-based
initiatives stemming from the PCA, which was
funded for three years by the Commonwealth. The
objective of the PCABC was to initiate a prototype
system for a Bioinformatics, Data Warehouse and
Biorepository Project, by developing: (1) a statewide data model for bioinformatics; (2) a statewide
repository of serum and tissue samples; (3) a data
model for biomarker data storage; and (4) a public
access website for disseminating research results
and bioinformatics tools. The seven institutions
comprising the Consortium have cooperated
closely to (1) explore the opportunity for sharing
clinical, genomic and other bioinformatics data on
patient samples in oncology, (2) for the purposes
of synergistically cooperating to share informatics
and other expertise, and (3) of developing collaborative research programs across cancer research
institutions in Pennsylvania. In many cases, clinical
specimens reside in individual departments in each
institution. These specimens are essentially
unknown to the comprehensive research community and thus underutilized. The Consortium’s
vision was to provide a virtual repository of clinical
samples located in various departments across the
state, supported by a common and well standardized informatics platform, in order to make them
available for research. In addition, one of our
primary goals was to facilitate the identification of
cancer-specific biomarkers and encourage collaborative research efforts among the participating
centers. Details of this “virtual biorepository” for
melanoma, breast and prostate cancers, tools for
data mining, and the imaging storage and analysis
software that make up the bioinformatics platform
will be discussed.
Methods
Participating institutions
The PCABC was created to explore the opportunity
for sharing clinical, genomic and other bioinformatics data on patient samples in oncology, for the
purpose of developing collaborative research
programs across cancer research institutions in
Cancer Informatics 2007:3
Multi-Institute Approach to Translational Cancer Research
Pennsylvania. It is hoped that this activity will
serve as a model for other regions, and be able to
include multiple institutions in an enhanced
research environment. The organization was
initially comprised of University of Pittsburgh
Cancer Institute, Fox Chase Cancer Center,
Abramson Cancer Center of the University of
Pennsylvania, Pennsylvania State Cancer Institute
at Milton S. Hershey Medical Center, Fox Chase
Cancer Center, Kimmel Cancer Center of Thomas
Jefferson University, and The Wistar Institute. The
Consortium has recently added new partners, that
include the Geisinger Health System, Windber
Research Institute, and Drexel University College
of Medicine, and it is considering further expansion
to include institutions from neighboring states.
The Consortium has access to cases and research
facilities from a variety of medical care settings
that include major cancer centers, academic
medical centers, research institutions, as well as
rural community-based healthcare networks. This
varied access allows the Consortium to encompass
the wide diversity of patients with cancer that
reflects the demographics of the State of Pennsylvania
and the surrounding areas and thereby expanding
beyond the patient populations of the Cancer
Centers’ component of the PCABC.
Organization of the consortium
The Consortium is governed by an Executive
Group made of PCABC institute leaders, who
oversee and guide Consortium’s development and
ongoing activities. Figure 1 depicts the organizational chart of the Consortium. The skill set of each
committee and types of roles that make up each
committee are summarized in Table 1. The Executive Group has delegated tasks to several subcommittees that included:
Disease/organ site sub-committees
Initially, a subset consisting of three target cancers:
breast and prostate cancers and melanoma were
selected to develop and implement a prototype
system for evaluation of the complex elements
necessary to enable the Consortium to succeed, as
well as leveraging each of the member institutions’
well established research programs. The extension
PCABC Organization Chart
PCABC
Executive Group
Working
Group
Common
Data
Element (CDE)
Sub-Committee
Institutional Review Board
(IRB)
Sub-Committee
Scientific Review
Committee
Disease Organ
Site
(DOS)
Sub-committees
IP/Tech Transfer
Sub-Committee
Each site organized by similar workflow/procedures:
Tissue Bank(ers) – does tissue collection and Quality Assurance
and Quality Control (QA/QC)
Tissue Bankers
Cancer Registrars
Data Managers
Cancer Registrar – works with tissue bankers and data managers
to ensure outcomes data collection. Acts as “honest broker”.
Oversees HIPAA, Quality Improvement (QI) intiatives.
Data Manager – does all QA/QC on entered data
Figure 1. Organization Chart. The PCABC organization chart.
Cancer Informatics 2007:3
257
Patel et al
Table 1. Skill set of committee members.
Summary of the required skill sets and roles that make up each of the committees of the PCABC organization.
Committee
Skill sets of committee members
Executive
Scientific Review Group
Cancer Center Directors and principle investigators of grant
Independent group of experts for each disease/organ category of interest
and a liaison member from the working group. Members include biostat
isticians, faculty, researchers, and physicians.
Intellectual Property (IP)/
Tech Transfer
Disease Organ Site
Institutional Review Broad (IRB)
Common Data Elements (CDE)
Working Group
General counsel, business development directors, technology transfer
directors, and faculty members
Leading experts of organ/disease of interest, faculty and researchers,
physicians.
Member of a IRB board, HIPAA regulatory specialist, tissue bank director,
researchers, physicians, data managers
Informaticians, molecular pathologists, microbiologists, oncology
informatics pathologists, genomics core directors, tissue bankers, cancer
registrars, data managers, pathologists, researchers and others who
bring a wealth of experience in collaborative data collection.
Tissue bankers, cancer registrars, data managers, pathologists,
researchers, informaticians, post-doctoral fellows, database
administrators and project coordinators.
to additional cancers, as prioritized by the Executive Group, will follow the successful implementation of the prototype system by all members of the
Consortium. In addition to storing the tissue
samples with the needed clinical annotations, the
Consortium aims to perform biomarker tests on
the list of biomarkers compiled and agreed to by
the Disease/Organ site sub-committee members.
The Master List of purposed biomarkers is
provided as supplemental file #1 with this manuscript and is a compilation of biomarkers suggested
by investigators from all member institutions. The
PCABC asked the research community to help the
Consortium to determine which markers would be
of interest strategically to compliment studies
already underway or to add to banks of information
already accumulated on a particular marker. For
each biomarker submission, the PCABC asked the
investigator to: 1) name the marker and primary
cellular function or pathway involved in its impact
on the neoplastic cell, 2) provide a short description
of the background work done on this marker and
why it would be important or at least worthwhile
to pursue it as a therapeutic and/or diagnostic
marker, and 3) designate the organ system site of
most interest for exploring this marker explored,
and the rationale behind its selection. The logic
was to ask these well-published disease site experts
and program leaders from each center to work with
the Consortium in identifying biomarkers that have
the most promise, with the intention of targeting
258
those biomarkers first and foremost. The Master
List of purposed biomarkers is a “work in progress,”
as each disease/organ sub-committee assembles to
ration and prioritize the biomarkers of interest. This
“Biomarker Master List” will be used as a guide
for all research proposals requesting biospecimens
from the Consortium. The Consortium does not
expect that each organ site and biomarker will be
studied by each institution. Instead, a scientific
review group will be assembled as part of the
committee structure of the PCABC, which will
decide which biomarkers are performed on which
tumor sites and by which partnering organizations.
If certain biomarkers are able to be tested at
multiple sites, the scientific review group along
with the executive committee will decide which
institution will be the preferred testing site and a
second testing site for validation. Experimental
results and biomarker data generated from the
tissues utilized from the Consortium biorepository
will be requested from the researchers at the
member sites for inclusion within the database as
described below.
Intellectual property (IP)/technology
transfer sub-committee
This group has the task of reviewing the issues
surrounding how to deal with any intellectual
property that develops as a result of the Consortium’s efforts. This group is identifying how the
Cancer Informatics 2007:3
Multi-Institute Approach to Translational Cancer Research
legislative language [6, 8] that dictates the grant
award affects IP. The committee will begin to
define potential IP and has devised a framework
inter-institutional agreement to deal with the
technology transfer needs of the Consortium.
Additionally, to protect each institution, the
Consortium has carefully studied the biomarker
list and their IP characteristics, to ensure that the
IP protected markers will be performed by the
institutions offering them and that there will be
agreements between the institutions to protect
both the patients’ privacy and confidentiality
when these biomarker assays are performed. This
committee includes general counsel, business
development directors, technology transfer
directors, and faculty with experience in this
important and complex area.
Institutional review board (IRB)
sub-committee
This sub-committee had the initial task of: a)
reviewing current practices within each center
regarding sharing of banked tissue, b) comparing
current patient informed consent samples, and
beginning to develop a common standardized
patient consent form and IRB protocol template,
that all centers will use when collecting samples
for analysis by the Consortium, and c) communicating with each center’s tissue banks to
determine what tissue types and the number of
samples that are sharable immediately. In
addition, it is the Consortium’s intention that in
subsequent years of the program, cross institutional IRB will be generated to review specific
uses of these tissues between institutions. For
example, if a group at the University of
Pittsburgh wanted to partner with the University
of Pennsylvania in a melanoma project, a
separate IRB that would address tissue specimens
collected at both institutions and the exact
nature of that research that would be performed
will be the subject of a subsequent IRB. Hence,
the initial proposed work will focus on allowing
the PCABC partners to bank and share
de-identified information through a central
database in the goal of providing this information for investigators across all member
institutions that are part of this Consortium. The
PCABC is fortunate to have on this committee
an IRB chairman, a HIPAA regulatory specialist
and a tissue bank director who can help direct
Cancer Informatics 2007:3
the dialogue and provide thorough guidance to
the Executive Group.
Common data elements (CDE)
sub-committee
This important sub-committee was one of the first
to be organized, in order to decide what information
on specimens and assays will be tracked. Standardization and compatibility will be paramount to
successful data sharing. To that end, the subcommittee has broken down the task into nine
modules: Patient demographics, family history,
clinical history, genomic information, patient
consent issues, pathology report, tissue sample
descriptors, outcomes and biomarkers. The intent
of this modularization is to enable generalizable
CDEs to be commonly developed across tissues/
organs/cancers and to facilitate the swapping in/out
of specialized modules as they are needed.
However, because one of the initial goals was to
enable the rapid implementation of a tissue and
data repository for the Consortium in support of
the biomarkers focus, the sub-committee focused
the CDE development on the modules that allowed
annotation of biospecimens. While utilizing and
learning from the experiences of several others
groups, the CDE sub-committee particularly took
into consideration the experiences of the Cooperative Prostate Cancer Tissue Resource [CPCTR,
see http://www.prostatetissues.org] [9, 10] as well
as established open source standards including the
Cancer Staging Manual[11] from the American
Joint Committee on Cancer, the Data Standards for
Cancer Registries from the North American Association of Central Cancer Registries [12], the
College of American Pathologist’s Cancer Checklist[13], and other biorepository specific CDEs that
were available through the National Cancer Institutes Center for Bioinformatics [14, 15]. This
committee is comprised of molecular pathologists,
microbiologists, oncology informatics pathologists, genomics core directors, and others who
bring a wealth of experience in collaborative data
collection.
Working group
PCABC’s Working Group infrastructure consists
of key personnel from each institution, which
includes data managers, cancer registrars, tissue
bankers, database administrators and project coordinators who are responsible for day-to-day
259
Patel et al
operations at individual sites as well as reporting
progress to the central project coordinator.
Data Managers from each participating site
act as the communicators and liaisons to the
project coordination center. Responsibilities
include managing the local site’s data entry and
the site’s own de-identification link back to the
patient identified data. A Data Manager is the
point person for PCABC to communicate with
regarding all clinical tissue and data collections
and interactions with tissue bank directors and
cancer registry managers from their respective
institution. Data Managers will eventually be
responsible for quality assurance standards
within their center as they relate to data entered
into the PCABC database.
Data Entry Personnel are any staff member who
has access to the PCABC database and has the
responsibility to review and/or enter data for their
institution. These individuals review and disseminate data from various internal sources. Data entry
personnel have the responsibility of entering the
data accurately into the PCABC database. Oftentimes, they work within the tissue bank or the tumor
registry department of the participating center, as
they have the training and experience necessary to
provide accurate and timely entry of critical data.
Each institution’s tissue bank(ers) is responsible for
collection and quality assurance and quality control
of the biospecimens shared with the Consortium.
The Cancer Registrars work with tissue bankers and
data managers, to ensure outcomes data collection,
and act as “honest brokers” (see below) as well as
assist in quality improvement initiatives with data
annotation processes.
Human subjects protection—the
honest broker concept
Each PCABC institute collects tissue and data
locally according to the local and institutional
guidelines and procedures. Tissue is stored locally
but data management is done centrally. Each
member organization has developed its own local
protocols, including consent language describing
its procedures to protect the confidentiality and
privacy of human subjects [16, 17] and has
obtained local/institutional IRB approval for all
PCABC activities.
The institutions that make up the Consortium
ensure protection of patient identity through “The
Honest Broker Concept”[18]. An “honest broker”
260
or “tissue bank trustee” acts as a well-defined
barrier between the clinical environment (in which
fully identified confidential patient information is
routinely exchanged as part of medical care) and
general research community (in which all information must be completely de-identified). In its
simplest form, the honest broker is not part of either
the clinical or research team and is the only person
or organization that can link research identifiers
and clinical identifiers. Although it differs between
each PCABC organization, provisions are in place
for the data managers, tumor registrars, tissue
bankers, or local database administrators at the
local PCABC sites to act as honest broker(s). Use
of the honest broker system allows each organization to control the de-identification process and
places responsibility and accountability for that
process. Personal and clinical identifiers (names,
medical record numbers, etc.) are limited to the
clinical space while research identifiers (i.e. “subject
12432” are never tied to the personal or clinical
identifies except in the honest broker’s code book.
This concept differs from anonymization, which
is a one way process of removing the linkage
between personal identifiers with research
identifiers.
This concept is implemented by designating at
least one tissue bank trustee from each institution
to the central PCABC coordination site. For
example, at one institution, a tumor registrar is
designated as the tissue bank trustee since tumor
registrars, by the nature of their job and by federal
mandate already have access to clinical information
on cancer patients and do not have access to the
results of research data for tissue bank samples. The
trustee is the only person who can link a patient with
the tissue bank number that identifies that patient.
The trustee system ensures that new clinical
outcome information can be added to a file identified
only by a code number, not a name. In rare but
possible circumstances, when critical research data
becomes available and it becomes medically necessary to inform the patient or his survivors, the
honest-broker system provides for an effective
mechanism to ensure that the critical information
reaches the interested party in a timely manner.
Central database design
The PCABC central database was modeled and
expanded from the work of the Cooperative
Prostate Cancer Tissue Resource (CPCTR) [9, 10].
Cancer Informatics 2007:3
Multi-Institute Approach to Translational Cancer Research
The system is relational database designed as a
virtual bio-repository that gathers information
on banked tissues and patients in research trials
—including clinical and molecular (gene and
protein data)—from each of the Consortium
member institutions. This data will be de-identified
at the local institutions and made available in a
central data warehouse for visualization and
query.
The workflow for entering data into the virtual
bio-repository has been as follows:
1. The local (physical) tissue bank identifies cases
appropriate for inclusion in the Consortium’s
virtual bio-repository (warehouse).
2. The local (physical) tissue bank pre-processes
data on these cases. The most important component of pre-processing is de-identification.
All de-identification occurs at the local banks.
No identifiable data is sent to the virtual biorepository (warehouse).
3. De-identified data are entered into the warehouse through a web site. The data entry web
site uses radio buttons, combo boxes and other
highly constrained data elements.
4. The local (physical) banks label each case with
a de-identified number. This number is used to
link the information in the warehouse to the
cases in the local banks. The linkage codes are
stored locally, using appropriate electronic and
physical safety measures. Only the local banks
have access to these linkage codes. The database
also keeps a system generated key that is tied
to the de-identified number that protects and
secures the data links as an additional security
measure.
5. The warehouse contains very minimal demographic data and complies with all HIPAA
requirements.
6. Access to the data entry application is controlled
by user name and password. Cases entered into
the virtual bio-repository are scanned for logical
errors (e.g. first recurrence before diagnosis
etc.).
The workflow for querying the warehouse is as
follows:
Initially, access is limited to members of the
Consortium using a username and password
system. The data manager at each facility is able
to provide user names and passwords for approved
researchers at that institution. The query tool
accesses the central database through a highly
Cancer Informatics 2007:3
constrained “click and point” interface. The data
in the warehouse will eventually will allow queries
on many of the approved data elements. However,
the specificity of the data returned will depend on
the user’s profile.
There are three user profiles as follows:
1. The Public Query tool is available to the general
public and is accessible through the Consortium’s web site [http://www.pcabc.upmc.edu].
The output display of a public query is the accrued number of cases, specimens and blocks
in the database that meet the criteria of the
query and general statistics on a limited number
of data elements. It is designed to allow interested investigators to see if there are enough
tissue specimens available through the PCABC
that meet their specific requirements.
2. Approved Investigator Query tool is password
protected tool and is distributed to those research
investigators who have approved research protocols within the PCABC or its member institutions. It allows users to refine and compile case
lists for their application and also to mine and
modify the default data views on the cases they
received biospecimens from the Consortium. It
shows all the annotations on the data associated
with each case through multiple pre-defined
views of the data set.
3. Data Manager Query tool is a password protected tool, available only for the internal
PCABC members. It is meant for data managers
to address QA issues regarding the data collected from a specific institute and generating
a tissue disbursement list. The main difference
between this viewer and the Approved Investigator Tool is that this tool allows the user to
identify the institution from which the cases
originated (i.e. UPCI, FCCC, PSU, etc.).
Application of use of material
Instructions on how to apply for specimens are
available by contacting the PCABC central project
coordinator. Should a researcher find tissue
samples that may be useful in ongoing or proposed
research from the PCABC query tool, the investigator is requested to first submit a letter of intent
(LOI) indicating their proposed study and tissue
requirements. The LOI is reviewed by the
Executive Group or the Consortium’s Scientific
Review Committee, consisting of an independent
261
Patel et al
group of experts for each disease/organ category.
Tissues are not released until after:
• IRB approval of the study at the researcher’s
institution—for patient safety of purposed
research.
• IRB approval of tissue collections at the tissue bank
institution—for patient safety of biospecimens
banked.
• Approval of the Consortium’s Scientific Review
Committee—to determine if the proposed
research has validity and justifies the use of
potentially valuable tissue resources from the
Consortium.
• Approval of the local (physical) tissue bank - As
“owner” and guardian of the tissue specimen
(the local banks may have their own review
committees or may wish to collaborate with the
researcher).
Results
The PCABC has become a model of how to
effectively partner across multiple and complex
research institutions, to achieve the common goal
of building a statewide or regional bioinformatics
network. Currently, there are over 80 individuals
who are members of this unique PCABC partnership, which is continuously looking to expand to
include new partners so that every region of the
state as well as neighboring states can contribute
and participate. This has created a network of
researchers, clinicians, bioinformaticians, cancer
registrars, program directors and external corporate
partners - all working together to accomplish a
common mission.
The various sub-committees have developed a
common IRB protocol template, common data
elements for standardizing data collections for
three organ sites, intellectual property/tech transfer
agreements, and material transfer agreements that
have been approved by each of the member institutions. This was the foundation work that has led
to the development of a centralized data warehouse
that has met each of the institutions’ IRB/HIPAA
standards. The data warehouse has web-based
interfaces for both data entry and query capabilities. The data entry is either done manually over
the internet or semi-automated batch files through
mapping of local data elements with PCABC
common data elements. A second, more important
part of the warehouse is its ability to associate
262
supplementary data sets, such as high throughput
molecular data or other files, such as whole slide
images of paraffin sections, by linking it with the
originating specimen (and patient). Although these
supplementary files are not fully integrated into
the biospecimen warehouse, hyperlinks to separate
databases or applications are provided for those
cases that are known to have any additional files.
The group has also developed a central web-based
tool that allows investigators to share their translational (genomics/proteomics) experiment data
on research evaluating potential biomarkers at a
central location on the Consortium’s website
[http://pcabc.upmc.edu/data/acctools]. In some
cases, access to raw experimental data would be
available to only internal PCABC members or
would require communication and collaboration
with the data originator. The many tools and capabilities of the PCABC data warehouse are highlighted in Figure 2.
The design of the central database is developed
to have normalized subset of standard clinical,
pathological, outcomes and molecular and other
biomarker descriptors, to enable Consortium
members to utilize the warehouse to substantially
complement their own local data repositories
while making tissue specimens optimally available for use in translational and clinical cancer
research.
Case accruals for the PCABC
biospecimen data warehouse
Access to well qualified tissue samples is fundamental to biomedical research. PCABC developed
this site for the exchange of de-identified information on consented tissue samples available to
Pennsylvanian and national researchers. This site
is updated each morning [http://www.pcabc.upmc.
edu/main.cfm?dis=pqe]. Figure 3 shows the
accrual rate of cases for each member institute by
the cancer type.
As of January 2007, the “virtual biorepository”
has over 60,000 individual annotated samples
from 12,734 cancer patients available for research
purposes. The Prostate Cancer Database currently
holds information on 7327 cases (including 9424
specimens and 46752 annotated blocks and blood
samples). The Breast Cancer Database currently
holds information on 3645 cases (including 4707
specimens and 10325 annotated blocks and blood
samples). The Melanoma Database currently
Cancer Informatics 2007:3
Multi-Institute Approach to Translational Cancer Research
holds information on 1762 cases (including 1591
specimens and 2953 annotated blocks and blood
samples). These numbers continue to grow, and
plans to add more tumor sites are in progress. The
accomplishments of the Consortium, since
funding of this project in 2002, are summarized
in Figure 4. A comparison of tissues banked with
the cancer incidence data (race and sex) from the
Pennsylvania Department of Health is described
by Figures 5–8.
Information and resources available
from the consortium
The PCABC website contains additional information
about the Consortium, including a frequently asked
questions (FAQ) section. Additionally, many of the
key resource documents have been provided as
supplemental files with this manuscript, such as:
• The latest version of the Biomarker Master List
(additional file #1)
• The Intellectual Property /Technology Transfer
Agreements (additional file #2)
• Biomarker Disclosure Form (additional
file #3)
• Template IRB Protocol for PCABC project
(additional file #4)
• Template IRB Protocol for the Honest Broker
system (additional file #5)
• Universal Consent Template (additional
file #6)
• Material Transfer Agreement Template
(additional file #7)
• The freely available CDEs developed for breast,
prostate, and melanoma by the Consortium can
be downloaded from its public database link on
the PCABC website. (additional files #8–10)
Alternatively, additional information can be
obtained by contacting the project coordinator
listed on the Consortium’s website [http://www.
pcabc.upmc.edu].
Discussion
Repository model
Each institution of the PCABC collects and stores
specimens locally within their own tissue banks,
while the data and information related to these
specimens are maintained in a centrally located
Quality Assurance/Quality Control
(QA/QC) Protocols
for data and biospecimens that are
shared to the Consortium
Upload Tools for Experimental Data sets
Data Storage Capability & Sharing
Mechanisms for translationalstudies and
other associated files.
Manual Data Entry Tools
using standardized CDEs
developed by PCABC
Data
Entry
Tools
Semi-automated Data Upload Tools
By mapping local legacy databases to
pre-formatted worksheets with
standardized CDEs
Data sets for biospecimens and clinical
history checked for logical errors and proper
vocabulary prior to acceptance into database.
Storage Capability
& Data Sharing
Mechanisms
for raw files of
experimental data
Data Managers
Query Tool
Query Tools
based on user
login
Public Query Tool
Approved Investigator
Query Tool
Experimental Data sets
Associated with specimens
Central
PCABC
Database
CDE Data
dictionary
Bioinformatics Tools:
Gene Expression Analysis Tool
-Access and categorize previously
analyzed microarray data with
Gene Ontology, OMIM, GenBank,
Locus Link,Homologene,
PubMed, etc
HTML links to associated
experimental data sets:
-Microarrray data
-Whole slide images
-Tissue microarray data
Figure 2. PCABC Database Tools. The tools and applications developed by the Consortium and data warehouse’s capabilities are
highlighted.
Cancer Informatics 2007:3
263
Patel et al
Cases accrual for each member
institute by the cancer type
3000
2500
2000
1500
Melanoma Cases
Breast Cases
Prostate Cases
1000
500
*CPCTR - VUC
*CPCTR - HOW
*CPCTR - MCW
*CPCTR - GWU
*CPCTR - NYU
Geisinger
Drexel
ACC, UPENN
KCC,TJU
FCCC
PSU
UPCI
0
*UPCI and UPENN are affiliates of PCABC and CPCTR; Partnership of both collaborative groups allowed centralizing available resources through the Consortium.
Source: PCABC Case Report March 2006
Figure 3. Case Accrual Rate. The total accrual rate of cases for each member institute by the cancer type.
bioinformatics and data management system that is
accessible by the members of the network. Although
the Consortium has been successful to date, the
implementation of this decentralized collection and
storage with a centralized data management model
has resulted in several limitations that are discussed
in the subsequent sections.
CDE issues
Creating the CDEs for the three initial disease types
was modeled from the work of the Cooperative
Prostate Cancer Tissue Resource. After their
successful implementation of the prostate CDEs,
the Consortium merely replicated the prostate
CDEs and modeled the same format for the breast
and melanoma CDEs within the data warehouse.
However, as data entry began for the two later
disease types, the difficulties of modeling them
264
after the prostate CDEs surfaced. In particular, the
prostate model allowed only one patient with one
primarily disease to be entered along with associated tissue accessions pertaining to the primary
disease; since once a prostate is removed, the
patient does not have another primary prostate
cancer. However, for other disease sites such as
breast and melanoma, data on additional primary
tumors of the same disease type cannot be accommodated within a single patient’s record using the
system developed. The central data warehouse was
a relational database, thus how data was entered
and queried was limited by the constraint put by
the design. The CDEs were arranged as:
• Case or Patient Identification number
o Consent Information
o Demographics and History
o Progression and Outcomes
Cancer Informatics 2007:3
Multi-Institute Approach to Translational Cancer Research
Summary of the accomplishments of the PCABC
Major milestones since initial funding of the project in 2002
Jun-03
Developed and now use a common
data entry tool over the web
(secure, Health Insurance Portability
and Accountability Act [HIPAA] compliant).
Aug-98
The Pennsylvania Cancer Alliance
(PCA) is organized.
Aug-02
Agreed to a solid data management
structure with clearly delineated
responsibilities and decision making.
Jul-03
Developed and now use a common
data query tool based on the
agreed common data elements.
Sep-05
Biomarker/Supplementary data annotation:
9 datasets of experiment results from
genomics/proteomics projects uploaded.
Jan-05
Jan-04
Jan-03
Jan-03
Agreed to a data warehouse model
for storing biospecimen, clinical,
biomarker data with capablities to
re-annotate experimental datasets.
Dec-04
Jul-05
Applied for and received
Applied for and received
$4.08M in federal funding
$500,000 pre-pilot project
“Cancer Biomedical Informatics Grid” caBIG
through Pittsburgh Life Science Greenhouse
[http://cabig.nci.nih.gov/]
New 7th and 8th partner:
leveraged from the established
synergy of the PCABC collaboration. Geisinger Health System and Windber Research Institute.
Apr-03
Developed new CDEs for
breast and melanoma
based on CPCTR’s
prostate model.
Jan-06
Mar-05
Over 375 whole slide images
on prostate cases linked to
originating specimen at patient level.
Jun-02
The PCABC received an initial
$5.6 million grant from the
Pennsylvania Department of Health
as part of a peer-reviewed competition
for Non-formula Tobacco Settlement funds.
Oct-05
Feb-04
CPCTR TMA #1
Establishment of partnership with
(299 prostate cases) linked to
NCI Cooperative Prostate Cancer
originating specimen at patient level.
Tissue Resource (6100 cases
May-04
Nov-05
and in 5th year of funding)
Establishment of partnership with
CPCTR TMA #2
[http://www.prostatetissues org]
NCI Cooperative Breast Cancer
(250 prostate cases) linked to
Mar-03
Tissue Resource (9000 cases
originating specimen at patient level.
Established procedures for
and in its 14th year of funding)
intellectual property/Technology transfer,
[http://cbctr.nci.nih.gov]
Feb-06
IRB protocols, universal consent,
Director’s Challenge dataset
patient privacy using Honest Brokers,
(62 prostate cases) linked to
and QA/QC of data and biospecimens.
originating specimen at patient level.
Figure 4. PCABC Timeline. Summary of the accomplishments of the PCABC since initial funding of the project in 2002.
o Clinical and Pathological Staging Information
• Treatment
• Biomarker data (i.e. PSA)
• Tissue Accession
o Accession Level Pathology data
o Block/Aliquot Level Pathology data
After initial implementation, several limitations
were faced when data entry was done on the
patients who had multiple primaries from the same
disease/organ site. For example, the current design
only allowed the input of staging information once
per patient record. To bypass this issue, separate
records with the patient identification number
assigned as 12345-A, 12345-B, etc. were entered
into the database for cases that had multiple
primary diseases from the same organ site.
However, this solution presents its own set of
issues. For instance, when a patient’s vital status
is presented as “dead”, the question of which of
the two primaries was the cause of death is raised.
Another issue is when updating this key information or other follow up data, multiple files of the
same patient would have to be updated for a single
piece of information. Furthermore, the ability to
search for cases with multiple primaries would not
Cancer Informatics 2007:3
be integrated, thus lessening the true value of such
cases for research studies.
IRB issues
Members of the IRB sub-committee addressed any
privacy issues that might be of concern at individual IRB committees. Developing a standard
IRB protocol that would be common at each institution for the entire project was a key step for the
Consortium, especially as it related to data sharing
in the PCABC biorepository data warehouse.
Several institutions encountered delays in getting
its IRB to approve the goals of the project. The
primary concerns of the IRB were 1) how and why
were data de-identified instead of annonymized?
2) Who has ownership of the tissue? 3) Who was
allowed the usage of tissues/data from the Consortium? And 4) how was it determined who was given
tissues/data from the Consortium?
In order to resolve these concerns, the project
coordinators conducted on-site meetings and
conference calls with individual site IRBs, to
address their concerns as well as educate and train
members of the Consortium on the role of cancer
registries, training of HIPAA guidelines, and imple265
Patel et al
Comparasion of Prostate Tissues Banked in PCABC
to Incidence reported to State DOH by Race#
100.0%
Percentage
80.0%
60.0%
PCABC
40.0%
STATE (PA)
20.0%
0.0%
White
Black
Other
Race
# Based on PCABC tissue accrual rates from all member institutions on June 2006. The PA State incidence data based on total number of cancer
cases by Race -Pennsylvania Residents, 1999-2003 ("PA Caner Incidence and Mortality Report 2003" by the PA Department of Health, published
July 2006).
Figure 5. Percentage of Prostate Tissues Banked compared to Incidence Reported in Pennsylvania.
Comparison of Prostate Tissues Banked in PCABC to Incidence reported to State DOH by Race.
mentation of the Honest Broker system [18]. It was
explained to the IRBs that it was critical to have
cases that were entered into the Consortium to be
de-identified so that follow-up information can be
added as well have the ability to re-annotate those
cases with experimental data sets when tissues
were used for research studies. It was also noted
that the ownership of the tissues is always retained
with the “tissue bank(s)” of each institution. Full
governance for all tissue and its associated data
was in place within the Consortium and distributed
to only those 0requests with all the appropriate
approvals mentioned above were obtained.
Furthermore, the creation of the multiple query
views based on the user’s profile was directly
related to the concerns of the IRB. An issue was
raised that although there was a governance place
on how tissue was requested, how was access to
the data warehouse limited to potential users.
Besides the privacy issues, the concern was that
having detail information on individual cases
266
would introduce case selection bias for potential
research studies.
Data annotation methods—manual
vs. automated
Originally, the partners envisioned primarily a manual
annotation methods by using web-based data entry
interfaces, but we quickly discovered this was too
time consuming as many centers had “legacy” databases that required researching and compiling
archived data. Developing standardized parsing and
semi-automated data entry methods with those institutions by adapting data mapping methods to created
flat files that allowed alternative methods of batch
uploads to the central data warehouse. However, the
ability to implement these semi-automated techniques
varied at each institution depended on the IT (information technology) support staff availability. Some
institutions had large sophisticated legacy databases
as well as access to other hospital base systems, such
Cancer Informatics 2007:3
Multi-Institute Approach to Translational Cancer Research
Comparasion of Breast Tissues Banked in PCABC to
Incidence reported to PA State DOH by Race^
90.0%
Percentage
80.0%
70.0%
60.0%
50.0%
PCABC
40.0%
STATE (PA)
30.0%
20.0%
10.0%
0.0%
White
Black
Other
Race
^ Based on PCABC tissue accrual rates from all member institutions on June 2006. The PA State incidence data based on total number of cancer
cases by Race - Pennsylvania Female Residents, 1999-2003 ("PA Caner Incidence and Mortality Report 2003" by the PA Department of Health,
published July 2006).
Figure 6. Percentage of Breast Tissues Banked compared to Incidence Reported in Pennsylvania.
Comparison of Breast Tissues Banked in PCABC to Incidence reported to State DOH by Race.
as tumor registry systems; those programmers and
database administrators were able to map the required
PCABC data elements with their internal data
sources. Some institutions had minimal informatics
infrastructure to support their homegrown databases
(i.e. Access), with limited or no domain knowledge
and expertise. These scenarios required the central
PCABC database administrators to educate those
institutions regarding the required format of the standardized export files. Of note, those institutions where
no previous tissue bank existed, the data entry tool
allowed for quick deployment of informatics tools
that not only helped with case accruals for the Consortium, but also for alternative ways of managing their
local tissue bank.
Issues of semantics
The legacy databases lack the ability to exchange
information and semantic interoperability, which
is the ability to understand and use the information
once it is received by other systems. Another
Cancer Informatics 2007:3
problems is that the panoply of ways that similar
or identical concepts or data are described by
different users even within the same institution. For
example, a data element called “grade of tumor”
can be collected using various formats (e.g. some
collect “grade-1, grade-2, and grade-3” vs. others
who collect “low grade, intermediate grade, and
high grade”). Such inconsistency in data descriptors makes it nearly impossible to aggregate and
manage even modest-sized data sets in order to be
able to ask basic queries. Moreover, these systems
are not uniform and flexible as well as incapable
of performing easy transfer of information and
unambiguous interpretation of the information
once it arrives into a central database.
However, our current database is based on such
a bioinformatics model that aids in developing and
conveying the semantic interoperability of the data
system by describing the common data elements
in the form of metadata or data descriptor (about
the content, quality, condition, and other
267
Patel et al
characteristics of the data) using controlled vocabulary and ontology, in order to make the data
understandable and sharable for end-users and
flexible for the system. Each common data element
is associated with an object or concept, attribute,
and valid value(s). For example, “patient age at
diagnosis” is the CDE that is made up of “patient”
(object), “age at diagnosis” (property) and the
representation (value domain) in “years”. Thus, in
order to share quality data into the PCABC central
database, the data collectors at each site need to
understand what exactly needs to be collected, as
has been agreed by the members of CDE subcommittee. Specifically for each of the approved
CDEs, the data collectors need to know: 1) the
fundamental definition of the data element (i.e.
date of diagnosis), 2) how that data element will
be collected (e.g. 11/2003 vs. Nov. 2003 vs. 11/03,
etc), 3) what are the consensus acceptable values
or codes are for the data element (e.g. precise date
of birth, not calculated from clinical records where
the “patient appears to be a well developed 75 year
old”), and 4) what the acceptable data format is for
inclusion into the central database (e.g. dates as
integers not character strings). Although the
concept of formalized metadata is fairly straight
forward, it has been rarely incorporated by clinical
and research groups building databases[19].
The CDEs developed by the Consortium are
ISO/IEC 11179 compliant (International Standards
Organization / the International Electro technical
Commission) [19,20] that means it defines a
number of fields and relationship for metadata
registries including detailed metamodel for
defining and registering items, of which the
primary component is a data element. The PCABC
uses CDE standards and metadata that are defined
by the consortium members of participating institutions. Furthermore, the Consortium members,
through their work in the caBIG initiative[21], are
working to enhance the database as a similar
object-relational model as PCABC as well as some
Comparasion of Melanoma Tissues Banked in PCABC
to Incidence reported to PA State DOH by Race*
100.0%
90.0%
86.3%
93.0%
Percentage
80.0%
70.0%
60.0%
PCABC
50.0%
STATE (PA)
40.0%
30.0%
20.0%
13.2%
6.6%
10.0%
0.5% 0.3%
0.0%
White
Black
Other
Race
*Based on PCABC tissue accrual rates from all member institutions on June 2006. The PA State incidence data based on total number of cancer
cases by Race and Sex-Pennsylvania Residents, 1999-2003 ("PA Caner Incidence and Mortality Report 2003" by the PA Department of Health,
published July 2006).
Figure 7. Percentage of Melanoma Tissues Banked compared to Incidence Reported in Pennsylvania.
Comparison of Breast Melanoma Banked in PCABC to Incidence reported to State DOH by Race.
268
Cancer Informatics 2007:3
Multi-Institute Approach to Translational Cancer Research
Comparasion of Melanoma Tissues Banked in PCABC
to Incidence reported to PA State DOH by Sex*
70.0%
Percentage
60.0%
50.0%
40.0%
PCABC
30.0%
STATE (PA)
20.0%
10.0%
0.0%
Male
Female
Sex
*Based on PCABC tissue accrual rates from all member institutions on June 2006. The PA State incidence data based on total number of cancer
cases by Race and Sex -Pennsylvania Residents, 1999-2003 ("PA Caner Incidence and Mortality Report 2003" by the PA Department of Health,
published July 2006).
Figure 8. Percentage of Melanoma Tissues Banked compared to Incidence Reported in Pennsylvania.
Comparison of Melanoma Tissues Banked in PCABC to Incidence reported to State DOH by Sex.
of the CDEs by using pre-defined controlled
vocabulary systems like those in NCI Thesaurus[22]
(NCI Thesaurus is a knowledgebase that contains
the working vocabulary used in data systems
covering clinical, translational and basic research
as well as administrative terminology) with semantically integrated and globally accepted valid
values for each data elements. PCABC has the
Oracle platform with internal object model,
whereas caBIG creates its own object model and
forms the data standard repository taking package,
class, attribute, and data type into consideration on
a Java template. This differs from the current
PCABC practices, in that, caBIG attempts to use
these “globally” accepted CDE standards and
metadata descriptors recommended by multiple
voluntary organizations and groups from the
research community that will allow all NCI-sponsored research to share data more easily. By taking
this new approach, greater flexibility will be given
to individual institutions with the current limitations
Cancer Informatics 2007:3
placed in data collection methods and work flow.
As another example, for the CDE “Patient race”,
the object is “patient”; property is “race” and the
valid value or representation of the metadata
(“Caucasian”, “African American”, “Asian”, etc).
If legacy databases from one institution uses
“African American” and another institute uses
“Black” as a valid value for the “race” field and
both their CDEs and metadata are properly mapped
to the central database, then when researchers
query for “African American” cases in the Consortium’s user interface, the result would display the
total cases (“African American” plus “black”)
available from both institutions, because both
terminologies are semantically integrated with the
accepted system. Hence, the overall advantage of
a distributed or federated model like caBIG over
a centralized model like PCABC include the shared
responsibilities of individual institutions for
services and implementation of the required standards and vocabulary that would foster data sharing
269
Patel et al
effortlessly. Furthermore, the centralized PCABC
architecture has a data sharing model inbuilt that
controls the various participating institutions. The
implementation of federated systems do not have
inbuilt data sharing models, but rather services that
run as external grid services, which connects with
local databases to establish such data sharing
mechanisms. The importance of semantic integration cannot be over emphasized.
Data integration of high throughput
datasets
Although more work needs to be done, the Consortium’s current database design has limited integration capabilities of supplementary data sets, such as
high throughput molecular data or other files, such
as whole slide images of paraffin sections, by linking
it with the originating specimen (and patient) [23–
25]. Unlike classical tissue annotation and clinical
data, high throughput techniques generate a large
amount of well formatted and semantically integrated data, typically for a relatively small number
of tissue specimens. These data require significant
computation analysis, usually using several different
approaches. The development of new tools for
analysis of molecular research data is presently an
area of intense research and focus of other major
initiatives in the vocabulary workspace domain of
caBIG infrastructure [21]. Data standards for high
throughput data are only now being developed and
there is no single standard that meets the needs of
all researchers. Development of a metadata repository may resolve this issue.
Tissue/data ownership issues
Numerous annotated tissue repositories already
exist at most large academic institutions and
cancer centers. In order to make them available
to a wider research community, much work needs
to be performed to educate those tissue banks
about the benefits of sharing their resources in
collaborative projects such as the PCABC. One
success of the PCABC was to take an inventory
of existing projects that possibly could partner
with the Consortium. This led to the partnership
with the CPCTR, with its data being shared in
the PCABC data warehouse. In addition, there
are multiple collections of similar disease/organ
types within a single organization that individuals
are either not aware of or cannot access.
Furthermore, many tissue bank-focused projects
270
do not consider the vast resources of paraffin
archives that are available for use, that is housed
in many academic pathology departments [26].
In order to resolve these issues, project leaders
from many developing and ongoing collaborative
projects [21, 27–36] need to provide incentives
to all parties, especially to all those tissue banks
that already exist, to collaborate and to build upon
what they have learned without “re-inventing”
much of the infrastructure.
Funding
The key feature to improve the management and
prevention of cancer is the novel ideas that come
from research, and the pace of research is
directly related to the availability of advanced
technology and top talent, together with the
infrastructure needed to financially support
them. A successful research effort therefore
requires a steady, reliable commitment of
support, and the tobacco settlement offers Pennsylvania an opportunity to make that crucial
long-term commitment and to measure the
results of its investment in lives saved.
On November 16, 1998, the Commonwealth of
Pennsylvania and 45 other states and territories
entered into an historic settlement with the tobacco
industry [5]. Under the terms of the Master Settlement Agreement, the tobacco industry will pay the
states $206 billion over 25 years. Pennsylvania’s
share of that settlement is valued at more than $11.2
billion [37]. This settlement presents Pennsylvania
with an opportunity to invest in research that will
reap benefits for all its citizens for generations. The
PCABC was awarded $5.5 million for a three year
project as one of the first unique collaborative
projects seeking to promote cancer research
throughout the state by utilizing the funds of the
tobacco settlement [38]. The Commonwealth’s
goal of funding such a project was to create a
bioinformatics infrastructure for a biorepository of
body fluids such as blood, serum, urine, etc. and
tissue samples which, in turn, put the Consortium
in a better position as a group to compete for additional funding to advance cancer and bioinformatics research. In addition, the Commonwealth
hoped to facilitate collaborations not only among
the researchers who are members of the Consortium, but also with outside collaborative partners
such as pharmaceutical and biotechnology companies, the Life Sciences Greenhouses [39], and
Cancer Informatics 2007:3
Multi-Institute Approach to Translational Cancer Research
researchers and institutions outside of the
consortium to energize economic growth throughout
the state.
According to a study produced by the Tripp
Umbach firm [40], since its start in 2001, the CURE
program has provided $298 million for medical
research in Pennsylvania using the tobacco money.
That has resulted in a $544 million boost to the
state’s economy—a nearly two-to-one return on
investment. More than 4,000 high paying jobs were
created or sustained in every sector of the economy
through CURE funding, and $32 million in new
tax revenue for the state was generated. In addition,
many members of the PCA successfully leveraged
the CURE funds to secure $138 million in new
federal medical-research funding as well as $500
thousand from the Pittsburgh Life Sciences Greenhouse [39] to add two new members to the Consortium. Figure 4 shows the additional funds generated
by the members of the PCABC.
In spite of the successes mentioned above, it
cannot be emphasized enough that one of the
fundamental issues related to building and
managing a biorepository is the large financial
commitment needed by various stakeholders
involved. In addition to the millions of dollars spent
by other international biorepository initiatives
[27–34], a recent report by the NCI Advisory
Bodies estimated over $50 million is spent annually
by the NCI on biorepository-related activities [41,
42]. Furthermore, a report by RAND corporation
[2] states that accurate determination of the actual
costs of collecting, processing, storing, and distributing tissue samples as well its supporting informatics infrastructure and operating costs need to
be determined as a “best practice” for cost recovery
to financially sustain future biorepositories. This
highlights the importance of standardizing and
sharing tools by all groups so that unnecessary
efforts are not duplicated, as well as, similar groups
can build upon the lessons learned by others. As
the Consortium prepares to open its biorepository
for requests by interested researchers, it is still
unclear whether it will be able to become selffunding through fees generated by specimen
requests.
Utilization and marketing
of the PCABC biorepository
As reveal above, a large financial commitment is
needed for a project like the PCABC to completely
Cancer Informatics 2007:3
be operational and self-sufficient. Leveraging the work
done by other collaborative groups [43, 44] that had
two member institutions of the PCABC was a key
factor for much of the success the Consortium
consummated with its limited funding. When
compared, each member of the NCI collaborative
groups received on average over $1 million/year
[45–47] as opposed to each of the initial seven PCABC
member institutions receiving approximately $262K/
year (PCABC total budget = $5.5mil/3years). The
members of the PCA do acknowledge that much work
still needs to be addressed in terms of marketing this
resource developed for the research community. Much
of the utilization of the biospecimens to date has been
limited to a handful of pilot studies between the
member institutions and numerous requests through
partnerships with collaborative tissue groups from the
NCI [43, 44]. Furthermore, the Consortium set its
priority on several issues that needed to be addressed
prior to the onset of utilization of the tissue resources.
These issues included (1) intellectual property issues
between the users and providers of the resources, (2)
IRB issues related to confidential and privacy sharing
of tissues and data across multiple institutions, and (3)
infrastructure issues dealing with a common bioinformatics platform. As described by other repositories
[3], marketing a biospecimen resource like the
PCABC to the broader research community will also
be a key factor in measuring its success for utilization
of the biospecimens. However, with substantial new
federal funding from the National Cancer Institute’s
cancer Biomedical Informatics Grid (caBIG, see
http://cabig.nci.nih.gov) initiatives [21], the members
of the PCABC are committed to sustaining and
continue to build the infrastructure developed by the
Consortium, and are confident that its investment will
provide an invaluable resource for large scale cancer
biomarker research, that will lead to the demonstration
of multiple new biomarkers as clinically important
guides to tumor classification, prognosis, response to
therapy, and/or clinical course of disease.
Conclusion
The technological achievements and the statewide informatics infrastructure that have been
established by the Consortium will enable robust
and efficient studies of biomarkers and their
relevance to the natural history and clinical
course of cancer. Studies resulting from the
creation of the Consortium may allow for better
classification of cancer types, more accurate
271
Patel et al
assessment of disease prognosis, a better ability
to identify the most appropriate individuals for
clinical trial participation, and better surrogate
markers of disease progression and/or response
to therapy. Of significance, the bioinformatics
infrastructure, data warehouse, and biorepository created through Consortium activities will
serve as a model as well as catalyst, and would
be made available for use by researchers and
institutions and other collaborative groups that
are presently not part of this research effort.
In summary, the Pennsylvania Cancer Alliances Bio-informatics, Data warehouse and
Bio-repository project has created a central
resource through which researchers can find
highly annotated tissue samples. The resource
will have no access at any time to patient identified data and tissue will not be made available
to researchers without the approval by IRB and
Scientific Review Committee. Furthermore, we
believe that the cancer centers of Pennsylvania
have the skills and experience necessary to
develop, build and implement a statewide
network of highly annotated tissue specimens
with relevant demographic, clinical, molecular,
and outcome data for biomedical research.
Based on the successful implementation of this
system, the Consortium hopes other groups
initiating similar projects will build upon the
lessons learned from the PCABC, for some of
the most important problems in biomedical as
well as clinical informatics, including medical
record de-identification, data mining tools and
complex data warehousing of biology data. By
integrating multimodal (pathological, clinical
and molecular) data in the annotated tissue
repository, creating query, visualizing techniques for these diverse data types and by
making this information available statewide, we
can help to provide much better selected and
characterized tissues for research. In the process,
we hope to foster new understanding of disease
evolution and progression that would eventually
help make Pennsylvania a leader in cancer
research and biomedical technology.
Acknowledgements
This work was supported in part by Pennsylvania Department of Health grant: PA DOH ME
01–740.
We acknowledge all the following contributors
to the development of the PCABC:
272
University of Pittsburgh Cancer Institute
Nancy Whelan, Lois Mathews, Sharon Winters,
Susan Urda, Heidi Gianella, Michelle Bisceglia,
Rajnish Gupta, Harpreet Singh, Songhui Li, Yimin
Nie, Vicky Chu, Sambit Mohanty, Dorothy Mann,
Linda Mignogna, Federico Monzon-Bordonaba,
Alice Katsur, John Kirkwood, Adam Brufsky,
Theresa Colecchia, and Carolyn Green.
Abramson Cancer Center of the
University of Pennsylvania
John Glick, Jesse Tigges, David Fenstermacher,
Timothy R. Rebbeck, Angela DeMichele, Barbara
Weber, DuPont Guerry and Elizabeth Poppert.
Fox Chase Cancer Center
Kate Haney, Steve Brusstar, John Malick, Kate
Haney, Arlene Capriotti, Andrew Balshem, Robert
G. Uzzo, Lori J. Goldstein, Stuart R. Lessin, Pat
Harsche and W.Thomas London.
Kimmel Cancer Center of Thomas Jefferson
University
Richard L. Davidson, Monica deBaca, Alejandra
Ruiz Orrico, Adler Hannes, Juan P. Palazzo, Adam
Dicker, Michael Mastrangelo and Katherine
Chou.
Pennsylvania State Cancer Institute
at Milton S. Hershey Medical
Thomas Loughran, Pam Whayland, Pat Swetland,
Philip Lazarus, Isom Harriet, Dan Beard, Thomas
Loughran, Alan J. Snyder, Witold B. Rybka, Daniel
Lorence, Allan Lipton, Harold A. Harvey, Gavin
Robertson, David Claxton and Richard Rauscher.
The Wistar Institute
Jenine Carlisle, Russel E. Kaufman, Donald Ewert,
Elizabeth O’Brien and Meryle Melnicoff.
Drexel University College of Medicine
Kenneth Blank, Tesfu Hailu, and Sokol Petushi.
Geisinger Health System
Glenn D. Steele, Jr., Sandy Buckley and Nicole
Hunter.
Windber Research Institute
Kim Yantus, Hai Hu, Chris Sheridan, Holly Rigby
and F. Nicholas Jacobs
Cancer Informatics 2007:3
Multi-Institute Approach to Translational Cancer Research
Pittsburgh Life Sciences Greenhouse
Pat Bronder, David Palmer and John H. Glick.
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
Becich MJ: 2000. The role of the pathologist as tissue refiner and
data miner: the impact of functional genomics on the modern pathology laboratory and the critical roles of pathology informatics and
bioinformatics. Mol. Diagn., 5(4):287–299.
Eiseman, E., Rand Corporation.: Case studies of existing human
tissue repositories: “best practices” for a biospecimen resource for
the genomic and proteomic era (http://www.rand.org/publications/
MG/MG120/). Santa Monica, CA: RAND; 2003.
Patel, A.A., Gilbertson, J.R., Parwani, A.V., Dhir, R., Datta, M.W.,
Gupta, R., Berman, J.J., Melamed, J., Kajdacsy-Balla, A., Orenstein
J. et al: 2006. An informatics model for tissue banks - Lessons learned
from the Cooperative Prostate Cancer Tissue Resource. B.M.C.
Cancer, 6(1):120.
Gilbertson, J.R., Gupta, R., Nie, Y., Patel, A.A., Becich, M.J.: 2004.
Automated clinical annotation of tissue bank specimens. Medinfo
11(Pt 1):607–610.
Master Settlement Agreement [http://www.attorneygeneral.gov/
uploadedFiles/Consumers/msa.pdf]
Pennsylvania’s Act 77 of 2001, the Tobacco Settlement Act [http://
www.legis.state.pa.us/WU01/LI/BI/BT/2001/0/HB0002P2378.
HTM]
Pennsylvania’s Health Research and CURE program [http://www.
dsf.health.state.pa.us/health/cwp/browse.asp?A=175&BMDRN=
2000&BCOB=0&C=38811]
Technology Transfer: Administration of the Bayh-Dole Act by Research
Universities [http://www.gao.gov/archive/1998/rc98126.pdf]
Patel, A.A., Kajdacsy-Balla, A., Berman, J.J., Bosland, M., Datta,
M.W., Dhir, R., Gilbertson, J., Melamed, J., Orenstein, J., Tai, K.F.
et al. 2005. The development of common data elements for a multiinstitute prostate cancer tissue bank: the Cooperative Prostate Cancer
Tissue Resource (CPCTR) experience. B.M.C. Cancer, 5:108.
Melamed, J., Datta, M.W., Becich, M.J., Orenstein, J.M., Dhir, R.,
Silver, S., Fidelia-Lambert, M., Kadjacsy-Balla, A., Macias, V., Patel,
A. et al: 2004. The cooperative prostate cancer tissue resource: a
specimen and data resource for cancer researchers. Clin. Cancer,
Res., 10(14):4614–4621.
Greene, F.L., American Joint Committee on Cancer., American
Cancer Society.: AJCC cancer staging manual, 6th edn. New York:
Springer-Verlag; 2002.
NAACCR Data Standards for Cancer Registries [http://www.naaccr.
org/index.asp?Col_SectionKey=7&Col_ContentID=122
The CAP Cancer protocols: Checklist for prostate gland [http://www.
cap.org/apps/docs/cancer_protocols/prostate04_pw.pdf
The NCICB’s Cancer Data Standards Repository (caDSR) [http://
ncicb.nci.nih.gov/NCICB/infrastructure/cacore_overview/cadsr]
Winget, M.D., Baron, J.A., Spitz, M.R., Brenner, D.E., Warzel, D.,
Kincaid, H., Thornquist, M., Feng, Z. 2003. Development of common
data elements: the experience of and recommendations from the early
detection research network. Int. J. Med. Inform., 70(1):41–48.
Department of Health and Human Services. 45 CFR (Code of Federal Regulations), 164.514(6)(2)(i). Standards for Privacy of Individually Identifiable Health Information (final) [http://www.hhs.
gov/ocr/regtext.html]
Health Insurance Portability and Accountability Act of 1996 [http://
aspe.hhs.gov/admnsimp/pl104191.htm]
Dhir, R., Patel, A.A., Winters, S., Grzybicki, D., Bisceglia, M., Swanson, D., Aamodt, R., Becich, M.J. 2006. A multi-disciplinary approach
to honest broker services for tissue banks and clinical data: a pragmatic and practical model. (In submission). CLin. Canc. Research.
Solbrig, H.R. 2000. Metadata and the reintegration of clinical information: ISO 11179. MD., Comput, 17(3):25–28.
Cancer Informatics 2007:3
[20] The ISO/IEC 11179 specification developed by the National Institute
of Standards and Technology (NIST) [http://isotc.iso.ch/livelink/
livelink/fetch/2000/2489/Ittf_Home/PubliclyAvailableStandards.
htm]
[21] The NCI’s Cancer Biomedical Informatics Grid (caBIG) [http://
cabig.nci.nih.gov/]
[22] NCI Thesaurus [http://nciterms.nci.nih.gov/NCIBrowser/Dictionary.
do]
[23] Berman, J.J., Datta, M., Kajdacsy-Balla, A., Melamed, J., Orenstein, J.,
Dobbin, K., Patel, A., Dhir, R. and Becich, M.J. 2004. The tissue microarray data exchange specification: implementation by the Cooperative
Prostate Cancer Tissue Resource. BMC Bioinformatics, 5(1):19.
[24] Li, S., Becich, M. and Gilbertson, J. 2004. Microarray Data Mining
Using Gene Ontology. Medinfo.778–782.
[25] Yu, Y.P., Landsittel, D., Jing, L., Nelson, J., Ren, B., Liu, L., McDonald,
C., Thomas, R., Dhir, R., Finkelstein, S. et al. 2004. Gene Expression
Alterations in Prostate Cancer Predicting Tumor Aggression and Preceding Development of Malignancy. J. Clin. Oncol., 22(14): 2790–2799.
[26] Patel, A.A., Gupta, D., Seligson, D., Hattab, E.M., Balis, U.J.,
Ulbright, T.M., Kohane, I.S., Berman, J.J., Gilbertson, J.R., Dry S.
et al. 2007. Availability and quality of paraffin blocks identified in
pathology archives: a multi-institutional study by the Shared Pathology Informatics Network (SPIN). B.M.C., Cancer, 7:37.
[27] The U.K. Biobank Project [http://www.biobank.ac.uk]
[28] Swedish National Biobank Program [www.biobanks.se]
[29] ATIP03.042. 2003. The Biobank Japan Project (http://www.atip.
org/public/atip.reports.03/atip03.042.pdf). In: Asian Technology
Information Program,.
[30] McCaffrey P. 2003. Iceland’s database tussle http://www.bio-itworld.
com/news/040103_report2255.html. Bio-IT World.
[31] Australasian Biospecimen Network (ABN) [http://www.abrn.net/]
[32] Thorn, H.C.D., Devereux, L., MacCallum, P. and Zeps N: 2005. A
more auspicious start for the Australasian Biospecimen Network
[Letter to the Editor]. J. Clin. Invest.
[33] The Singapore Tissue Network [http://www.stn.org]
[34] Spinney, L. 2003. U.K. launches tumor bank to match maligned
Biobank. Nat. Med., 9(5):491.
[35] The National Biospecimen Network (NBN) blueprint
[http://biospecimens.cancer.gov/nbn/blueprint.asp]
[36] VonEschenbach, A.C., Buetow K. 2006. Cancer Informatics Vision:
caBIG. Cancer Informatics, 2:22–24.
[37] Summary of Key Points in the Master Settlement Agreement [http://
www.attorneygeneral.gov/consumers.aspx?id=683]
[38] CURE Awarded Grants: Health Research Nonformula Grant Awards
for SFY 2001–02 [http://www.dsf.health.state.pa.us/health/lib/health/
cure/2001_nonformula_grants_sfy_01–02.pdf]
[39] The Pittsburgh Life Sciences Greenhouse [http://www.pittsburghlifesciences.com]
[40] George J. 2006. Study details tobacco settlement’s boon to Pennsylvania. In: Philadelphia Business Journal.
[41] National Cancer Advisory Board meeting minutes, November 2004
[http://deainfo.nci.nih.gov/advisory/ncab/132_1104/30nov04mins.pdf]
[42] National Cancer Institute’s Board of Scientific Advisors meeting
minutes, November 2004 [http://deainfo.nci.nih.gov/advisory/bsa/
bsa1104/08nov04mins.htm]
[43] C o o p e r a t i v e P r o s t a t e C a n c e r T i s s u e R e s o u r c e
[http://www.prostatetissues.org]
[44] T h e C o o p e r a t i v e B r e a s t C a n c e r T i s s u e R e s o u r c e
[http://www-cbctr.ims.nci.nih.gov/]
[45] Cooperative prostate cancer tissue resource: Release Date April
29,1999, RFA CA-99–012, National Cancer Institute [http://grants1.
nih.gov/grants/guide/rfa-files/RFA-CA-99–012.html]
[46] Cooperative human tissue network: Release Date March 12, 2000.
RFA CA-01-009, National Cancer Institute [http://grants.nih.gov/
grants/guide/rfa-files/RFA-CA-01–009.html]
[47] Cooperative breast cancer tissue registry: Release Date February
19,1993, RFA CA-93-19, National Cancer Institute [http://grants2.
nih.gov/grants/guide/rfa-files/RFA-CA-93–019.html]
273
Patel et al
Supplemental Files
• The Master List of purposed Biomarkers
(additional file #1)
• The Intellectual Property /Technology Transfer Agreements (additional file #2)
• Biomarker Disclosure Form (additional
file #3)
• Template IRB Protocol for PCABC project
(additional file #4)
• Template IRB Protocol for the Honest Broker
system (additional file #5)
• Universal Consent Template (additional
file #6)
• Material Transfer Agreement Template
(additional file #7)
• The freely available CDEs developed for breast,
prostate, and melanoma by the Consortium can
be downloaded from its public database link on
the PCABC website.
o
Prostate CDEs (additional files #8)
o
Breast CDEs (additional files #9)
o
Melanoma CDEs (additional files #10)
274
Cancer Informatics 2007:3
Download